Mutation-based evolutionary algorithms and their applications to optimization of antennas in layered media

We first review the details of two modified evolutionary programming (EP) algorithms, meta-EP and fast evolutionary programming (FEP). These algorithms combine the mutation-based evolution of EP with the self-adaptation of the strategy parameters in ES. The algorithms are then applied to the challenging constrained optimization problem of the printed Yagi-like array of stacked dipoles. Due to the presence of surface-waves and strong mutual coupling effects, the objective function relating the gain to antenna parameters is highly non-linear and possess multiple local maxima. Examples of optimization of the Yagi-like array printed in isotropic as well as uniaxial dielectric layers are given. For the latter case, it is possible to obtain high gain radiation at two different elevation angles in the E and H planes.